Update to 9.42
Upstream changes:
9.42 2025-10-01
- Un-deprecated the spurt method in Mojo::File, it is now an alternative to spew.
- Removed experimental status from top-level await support in Mojo::Promise.
- Removed experimental status from encrypted session cookie support.
- Removed experimental status from persistent cookie support.
- Removed experimental status from samesite cookie support.
- Removed experimental status from colourful log messages.
- Removed experimental status from freeze option in Mojo::IOLoop.
- Removed experimental status from check and raise functions in Mojo::Exception.
- Fixed Cpanel::JSON::XS compatibility issues. (ilmari)
- Fixed async/await memory leak in Mojo::Promise. (TFBW)
9.41 2025-07-03
- Added EXPERIMENTAL support for Server-Sent Events.
- Added EXPERIMENTAL module Mojo::SSE.
- Added EXPERIMENTAL sse attribute to Test::Mojo.
- Added EXPERIMENTAL get_sse_ok, post_sse_ok, sse_finish_ok, sse_finished_ok, sse_ok, sse_id_is, sse_id_isnt,
[9 lines not shown]
Update to 2.002006
Upstream changes:
2.002006 2025-08-16
- Fix a parse-time memory leak with undef or default initializers
('fun ($x = undef)' or 'fun ($ =)'). This leak mostly manifests as
noise from memory checkers because non-leaked ops aren't freed
anyway while perl is still running.
- Update bugtracker info in README.
2.002005 2025-01-19
- When debugging (perl -d), skip over invisible (generated) parameter
initialization code even if single-step mode is active.
- Move repository and bugtracker to <https://codeberg.org>.
Update to 1.26
Upstream changes:
1.26 2021-02-16
- Parse fortran compilers with complicated names - thanks @d-lamb
- Get compiler from %ENV - thanks @d-lamb
v1.25 released Jan 27th 2020
---------------------------
* Fix of version number comparison in F77.pm link_gnufortran_compiler routine
Update to 0.8002
Upstream changes:
0.8002
- fix UNC path bug - https://rt.cpan.org/Ticket/Display.html?id=102436
- fix get_makefile_vars losing path components if space-quoted and >1 dep,
which was breaking CPAN module "true" which depends on two modules
Update to 6.037
Upstream changes:
6.037 2026-01-25 16:18:09-05:00 America/New_York
- avoid warning on undef input to prompts
- when prompting for PAUSE password, suggest that leaving it blank is
okay
- tweak autogenerated files to be less likely to be misidentified as
Prolog by Vim (ugh!) (thanks, Graham Knop!)
- when listing dependencies but skipping anything satisfied by core
perl version X, do not list perl itself (if version X is good enough)
(thanks, Leon Timmermans)
6.036 2025-11-09 18:07:15-05:00 America/New_York
- re-fix the re-exporter format_author_deps (thanks, Graham Knop!)
- fix Pragmas so it enables strict (!!!) (thanks, Graham Knop!)
6.035 2025-11-09 11:26:39-05:00 America/Toronto
- re-rename subroutine in Dist::Zilla::Util::AuthorDeps; these were not
meant to be relied upon, but Hyrum's Law applies
6.034 2025-11-07 09:20:18-05:00 America/Toronto
[9 lines not shown]
dwc_gmac.c: fix spelling in a comment
(Evidently there are several NIC drivers that have variations of the
"randomly generate a MAC address" approach already.)
rtl8169.c: generate MAC address for hardware without an EEPROM
Borrow the equivalent change from if_ure.c (r. 1.33). These re(4)
devices otherwise end up with an all zero MAC. Found (at least) on the
NanoPi R4S (non-"Enterprise" version, the other being distinguished by
including an EEPROM to set a consistent MAC address).
(Given this code is in two places now, it may make more sense to
centralize this if it grows any further.)
